Age is the primary risk factor for Alzheimer's disease (AD). Human studies that focus on the connections between transcriptomics and epigenomics of aging and their relationship to AD hold great promise for unraveling the molecular basis of AD and providing novel therapeutic targets. Transcriptomic and epigenomic studies reported to date have been hampered by a number of methodologic issues, especially the use of heterogeneous cell mixtures. Using transcriptomic profiles in purified monocytes from 1,263 participants of the Multi-Ethnic Study of Atherosclerosis (MESA), we reported a transcriptional network of co-expressed oxidative phosphorylation (OXPHOS) genes that decline with age. Our following transcriptomic analysis demonstrated that this OXPHOS network of 21 genes (FDRs<0.05) were positively associated with cognitive function. These human data, combined with our non-human primate data correlating mitochondrial function of monocytes and frontal cortex tissue and recent data in transgenic mice showing a causal role of mitochondrial dysfunction in AD, suggest that monocyte transcriptional profiles may reflect brain bioenergetic dysfunctions linking age to AD. Given emerging evidence in mice that monocytes can infiltrate the brain and take over immune surveillance, our data also suggest that altered monocyte function may affect AD. Nevertheless, our cross- sectional human data cannot determine whether genomic alteration of OXPHOS is a cause or consequence of AD. The goal of the proposed study is to determine the impact of cell specific gene networks, especially aging- related networks such as OXPHOS, on the development of AD through an integrated analysis of genomic, epigenomic and transcriptomic data in a longitudinal community-based study. The deeply phenotyped MESA cohort, with unique existing epigenomic, transcriptomic and cognitive data collected at Exam 5 (2010-11), offers an ideal study population. We propose to repeat the epigenomic, transcriptomic and cognitive assessment at Exam 6 (2016-17) and perform the cognitive battery of the Uniform Data Set to enable ascertainment of mild cognitive impairment (MCI) and Alzheimer's dementia at Exam 6 and three years later in a subset of MESA cohort (N=1,200) to achieve the following specific aims: 1) To determine whether aging- related changes in transcriptomic/epigenomic profiles predict cognitive decline over a 6-year follow-up; 2) To determine whether aging-related changes in transcriptomic/epigenomic profiles predict development of AD over a three-year follow-up; and 3) To determine whether differences in mitochondrial activity and content, which would be predicted from the OXPHOS alterations, relate to development of AD. This large, prospective multi-omic study will take a novel approach of focusing on aging-related gene networks in human homogeneous cells to identify new biomarkers for AD development, a crucial step forward in understanding cell-specific mechanisms and providing clues to early intervention.